Nancy Wanjiru

Key Skills

Software

Lionbridge

Top Subject Matter

Computational Linguistics

English NLP

Llms Domain Expertise

Top Data Types

Text

Audio

Top Task Types

Prompt Response Writing SFT

Classification

Freelancer Overview

Senior Language Specialist – AI Trainer, Outlier. Brings 3+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal, Proprietary Tooling, and Lionbridge. Education includes Master of Arts, University of Nairobi (2018) and Bachelor of Arts, Kenyatta University (2016). AI-training focus includes data types such as Text and labeling workflows including Prompt + Response Writing (SFT), Classification, and Translation.

Labeling Experience

Senior Language Specialist – AI Trainer, Outlier

TextPrompt Response Writing SFT

As a Senior Language Specialist—AI Trainer at Outlier, I designed and executed adversarial prompt scenarios to test advanced model capabilities. My main tasks included annotating training samples with metacognitive reasoning and building error-trace taxonomies for English language model evaluation. My role required me to work closely with cross-functional teams to refine prompt-engineering pipelines and evaluation metrics. • Designed 200+ syntactic and pragmatic adversarial prompt-test cases to expose LLM weaknesses • Annotated 3,500+ English training samples with explicit, context-rich rationales • Created a reproducible taxonomy cataloging 18 English model failure modes • Collaborated on prompt refinement, improving model accuracy and reducing hallucination rates

As a Senior Language Specialist—AI Trainer at Outlier, I designed and executed adversarial prompt scenarios to test advanced model capabilities. My main tasks included annotating training samples with metacognitive reasoning and building error-trace taxonomies for English language model evaluation. My role required me to work closely with cross-functional teams to refine prompt-engineering pipelines and evaluation metrics. • Designed 200+ syntactic and pragmatic adversarial prompt-test cases to expose LLM weaknesses • Annotated 3,500+ English training samples with explicit, context-rich rationales • Created a reproducible taxonomy cataloging 18 English model failure modes • Collaborated on prompt refinement, improving model accuracy and reducing hallucination rates

2024 - Present

Computational Linguist, Lionbridge

LionbridgeTextClassification

As a Computational Linguist at Lionbridge, I led multiple English NLP data labeling cycles for enterprise machine translation clients. My work centered on validating, classifying, and annotating over 10,000 translated text segments for quality metrics. I also developed team training modules to increase annotation quality and reliability. • Labeled and validated English text segments for fluency, adequacy, and cultural accuracy • Integrated morphosyntactic and pragmatic evaluation rubrics for annotation consistency • Increased inter-annotator agreement by 19% with a new English labeling framework • Produced technical evaluation reports and pipeline adjustment recommendations

As a Computational Linguist at Lionbridge, I led multiple English NLP data labeling cycles for enterprise machine translation clients. My work centered on validating, classifying, and annotating over 10,000 translated text segments for quality metrics. I also developed team training modules to increase annotation quality and reliability. • Labeled and validated English text segments for fluency, adequacy, and cultural accuracy • Integrated morphosyntactic and pragmatic evaluation rubrics for annotation consistency • Increased inter-annotator agreement by 19% with a new English labeling framework • Produced technical evaluation reports and pipeline adjustment recommendations

2022 - 2023

Language & Content Specialist, Safaricom PLC

Text

As a Language & Content Specialist at Safaricom PLC, I managed English data labeling and localization for the M-Pesa chatbot and IVR AI systems. My role involved comparative linguistic audits of AI-generated responses and maintaining an internal style and error guide. This work supported high-quality, culturally sensitive English dialogue flows within AI-powered customer solutions. • Labeled and localized English conversational flows for AI chatbot and IVR applications • Audited model-generated English text for tense, politeness, and grammar errors • Documented systematic annotation errors and taxonomy updates for internal use • Managed and updated style guides and terminology for team-wide linguistic consistency

As a Language & Content Specialist at Safaricom PLC, I managed English data labeling and localization for the M-Pesa chatbot and IVR AI systems. My role involved comparative linguistic audits of AI-generated responses and maintaining an internal style and error guide. This work supported high-quality, culturally sensitive English dialogue flows within AI-powered customer solutions. • Labeled and localized English conversational flows for AI chatbot and IVR applications • Audited model-generated English text for tense, politeness, and grammar errors • Documented systematic annotation errors and taxonomy updates for internal use • Managed and updated style guides and terminology for team-wide linguistic consistency

2020 - 2022

Research Assistant—Linguistics & Digital Humanities, Strathmore University

TextClassification

As a Research Assistant in Linguistics & Digital Humanities at Strathmore University, I performed large-scale English corpus annotation across multiple sociolects. This included supervised tagging of 15,000+ tokens and error taxonomy development for computational linguistics experiments. My mentoring of junior researchers helped maintain annotation quality and reliability. • Annotated large English corpora for linguistic research and model training • Developed and refined error-taxonomy for English part-of-speech tagging • Assisted in prototype development for annotated POS tagger • Trained and supervised junior annotators on annotation standards and agreement metrics

As a Research Assistant in Linguistics & Digital Humanities at Strathmore University, I performed large-scale English corpus annotation across multiple sociolects. This included supervised tagging of 15,000+ tokens and error taxonomy development for computational linguistics experiments. My mentoring of junior researchers helped maintain annotation quality and reliability. • Annotated large English corpora for linguistic research and model training • Developed and refined error-taxonomy for English part-of-speech tagging • Assisted in prototype development for annotated POS tagger • Trained and supervised junior annotators on annotation standards and agreement metrics

2018 - 2020

Master of Arts Capstone Project—Linguistics, University of Nairobi

TextClassification

For my Master of Arts in Linguistics at University of Nairobi, I built an annotated error corpus for English machine translation. I analyzed morphophonological errors and proposed rule-based post-editing solutions using an 8,000-segment dataset. This academic project centered on annotation and classification of translation errors for NLP applications. • Constructed and annotated an English error corpus for machine translation research • Classified errors for model evaluation using linguistic and computational metrics • Developed a rule-augmented post-editing pipeline • Applied annotation results to shape error analysis and model improvement strategies

For my Master of Arts in Linguistics at University of Nairobi, I built an annotated error corpus for English machine translation. I analyzed morphophonological errors and proposed rule-based post-editing solutions using an 8,000-segment dataset. This academic project centered on annotation and classification of translation errors for NLP applications. • Constructed and annotated an English error corpus for machine translation research • Classified errors for model evaluation using linguistic and computational metrics • Developed a rule-augmented post-editing pipeline • Applied annotation results to shape error analysis and model improvement strategies

2016 - 2018

Show more

Education

Work History